Thrombosis Research 96 (1999) 65–71
REGULAR ARTICLE
Evaluation of the Automated
Coagulation Analyzer SYSMEX CA 6000
Peter Quehenberger, Stylianos Kapiotis, Sylvia Handler, Katharina Ruzicka and Wolfgang Speiser
From the Clinical Institute of Medical and Chemical
Laboratory Diagnostics, University of Vienna, A-1090 Vienna, Austria.
(Received 16 September 1998 by I. Pabinge; revised/accepted 28 March 1999)
Abstract
In the present study the coagulation analyzer SYS-
MEX CA 6000 (TOA Medical Electronics Co.,
Kobe, Japan), an analyzer equipped with a photo-
optical clot detection unit and a cap-piercing sys-
tem, was evaluated with respect to its technical
characteristics in the determination of standard co-
agulation tests (prothrombin time, activated partial
thromboplastin time, thrombin time, fibrinogen,
and antithrombin) and in the determination of co-
agulation single factor activities. In the normal and
in the pathological range the intraassay coefficients
of variation (CV) and interassay CV for most pa-
rameters were below 5% (exceptions: intraassay
CV 5.4% for prolonged thrombin time; intraassay
CV 9.26% and interassay CV 10.7% for decreased
antithrombin; interassay CV 5.62% for fibrinogen
in the normal range, intraassay CV 10.1% for fi-
brinogen greater than 7.0 g/L; intraassay CV 6.36%
and interassay CV 11.7% for decreased fibrinogen;
interassay CV 11.6% for prolonged activated par-
tial thromboplastin time; interassay CV 6.12% for
decreased factor VII). Interference studies with
lipemic, icteric, and hemolytic samples showed just
minor influences of these abnormal sample charac-
Abbreviations: PT, prothrombin time; aPTT, activated partial
thromboplastin time; TT, thrombin time; AT, antithrombin; CV,
coefficient of variation; f.c., final concentration.
Corresponding author: W. Speiser, Clinical Institute of Medical
and Chemical Laboratory Diagnostics, General Hospital Vienna,
Währinger Gürtel 18-20, A-1090 Vienna, Austria. Tel.: 143 (1)
40400 5369; Fax: 143 (1) 4036688; E-mail: ,wolfgang.speiser@
univie.ac.at..
0049-3848/99 $–see front matter  1999 Elsevier Science Ltd. All rights reserved.
PII S0049-3848(99)00069-9
teristics on prothrombin time, activated partial
thromboplastin time, fibrinogen, and antithrombin
measurements when compared to the results ob-
tained by using mechanical clot detection (STA,
Stago Diagnostica, Asnieres-Sur-Seine, France).
No carryover was detected in alternating measure-
ments of heparinized (3 U/mL unfractionated hep-
arin) and normal plasma samples. Measurement of
the activities of clotting factors V, VII, VIII, and
IX showed a good correlation (r50.993 to r50.977)
between SYSMEX CA 6000 and STA. Our results
demonstrate that using SYSMEX CA 6000 ana-
lyzer basal routine coagulation testing as well as
specialized tests for single factor activities can be
performed with satisfactory precision; in particular,
the cap-piercing system has no negative effect on
the performance of the analyzer.  1999 Elsevier
Science Ltd. All rights reserved.
Key Words: Evaluation; Coagulation analyzer; Sysmex
CA 6000
D
ue to increasing demands for coagulation
testing, automated coagulation analyzers
have been developed for screening large
numbers of samples in coagulation laboratories.
Automated coagulation analyzers are expected to
be labor saving, quick, and safe, and should ensure
accuracy and precision. In the last years new types
of coagulation analyzers were developed that are
capable of performing not only standard coagula-
tion tests such as prothrombin time (PT), activated
partial thromboplastin time (aPTT), or fibrinogen,
66 P. Quehenberger et al./Thrombosis Research 96 (1999) 65–71
but also more sophisticated analysis such as mea-
surement of single coagulation factors activities.
Furthermore, these analyzers are usually equipped
with a photometric detection unit for chromogenic
substrate assays.
In the present study we evaluated the new coagu-
lation analyzer SYSMEX CA 6000 (TOA Medical
Electronics Co., Kobe, Japan) [1]; the most striking
features of the analyzer are cap-piercing technique,
fully automated sampling, positive identification of
samples by bar-code detection, reagent cooling to
158C, and the option to integrate the machine into
a fully automated “advanced coagulation analysis
system.” Coagulation tests are performed using a
photo-optical clot detection unit that is also suit-
able for chromogenic substrate analysis [2]. It was
our aim to evaluate the technical characteristics of
SYSMEX CA 6000 to obtain new information (1)
on the effect of cap piercing on the performance
of the analyzer; (2) on the comparability of values
obtained by using photo-optical clot detection with
those measured by using mechanical clot detection,
in particular in samples with abnormal optical char-
acteristics (lipemic, icteric, and hemolytic samples);
and (3) on the precision of automated single factor
coagulation activity measurements.
1. Materials and Methods
1.1. Characteristics of the SYSMEX CA 6000
The SYSMEX CA 6000 (TOA Medical Electronics
Co., Kobe, Japan) is a fully automated coagulation
analyzer equipped with a cap-piercing needle for
the transfer of plasma from the closed primary
tubes to secondary tubes located on a sample rotor
within the analyzer, from which samples are further
transferred by a separate pipettor. Primary tubes
are proceeded to the machine with the help of
sample racks, whereby a maximum of 50 samples
can be loaded at once using automatic sample iden-
tification by bar-code. Nine positions on the sample
table can be used for statim samples. Using the
current software (version 00-23C), 20 different tests
with five programmable steps for each test can be
adapted on the machine. Two pipettors are avail-
able for reagent distribution—one of these is in-
tended to pipette thrombin. There is a cooled re-
agent holder table (158C) with 28 positions for
reagents and two fixed positions for rinse solution.
1.2. Patients, Volunteers, and Blood Sampling
Citrated blood (Na-citrate, final concentration
(f.c.) 0.013 M) was taken into siliconized glass tubes
(Vacutainer Becton Dickinson, Meylan Cedex,
France). Platelet-poor plasma was prepared by
centrifugation at 30003g for 20 minutes. Plasma
was then assayed immediately or frozen at 2708C
until use. Determinations were performed on sam-
ples from patients whose specimens were sent to
the routine coagulation laboratory of our hospital.
For definition of normal ranges and estimation of
analysis time, samples from 40 apparently healthy
volunteers (20 females, 20 males; 25–58 years of
age, median 39 years) were used. The following
normal ranges (values within the 5th to 95th per-
centile) were determined: PT, 11.8–13.5 seconds;
aPTT, 30–41 seconds; thrombin time (TT), ,19 sec-
onds; fibrinogen, 2.0–3.5 g/L; antithrombin (AT),
80–100%; factor V, 74–136%; factor VII, 62–144%;
factor VIII, 60–180%; factor IX, 71–150%.
1.3. Reagents and Laboratory Methods
PT was performed by mixing 50 mL of plasma (pre-
incubated for 180 seconds at 378C) with 100 mL
PT reagent (Dade Thromboplastin IS; Dade Diag-
nostics AG, Duedingen, Switzerland, International
Sensitivity Index: 1.35) at 378C. aPTT was per-
formed by first mixing 50 mL Dade Actin-FS aPTT
reagent (Dade Diagnostics AG) with 50 mL
plasma; after an incubation period of 120 seconds
at 378C, 50 mL 0.025 M CaCl2 was added. For mea-
suring TT, 100 mL plasma were incubated for 240
seconds at 378C; 100 mL Dade Thromboclotin (f.c.
3.3 NIH units of thrombin, Dade Diagnostics AG)
were then added. Fibrinogen was measured by mix-
ing 100 mL of a 1:20 plasma dilution (Owren’s dilu-
ent buffer, preincubated for 240 seconds at 378C)
with 80 mL of fibrinogen reagent (Immuno AG,
Vienna, Austria); Fibrinogen was calibrated using a
1:10, 1:20, 1:40, and 1:80 dilution of Immuno Normal
Plasma (Immuno AG, Vienna, Austria) with Ow-
ren’s diluent buffer. Antithrombin was measured
using the assay kit STA Antithrombin (Stago Diag-
nostica, Asnieres-Sur-Seine, France). First, 100 mL
thrombin (f.c. 0.33 NIH units) were added to
 P. Quehenberger et al./Thrombosis Research 96 (1999) 65–71
100 mL of a 1:20 dilution of the sample; after 60
seconds of incubation at 378C, 100 mL of the chro-
mogenic substrate solution were added. For cali-
bration of AT, pooled normal plasma was diluted
1:20, 1:30, and 1:60 with Owren’s diluent buffer.
The test configurations of PT, aPTT, TT, fibrino-
gen, and AT were similar on SYSMEX CA 6000
and on STA analyzer (Stago, Asnieres-Sur-Seine,
France). The activities of the coagulation factors
V and VII were measured by mixing 50 mL of 1:10,
1:20, and 1:40 plasma dilutions (Owren’s diluent
buffer) with 50 mL factor V (Behringwerke, Mar-
burg, Germany) or factor VII-deficient plasma
(Dade Diagnostics AG). The mixtures were then
incubated for 180 seconds (on STA for 60 seconds)
at 378C and thereafter 100 mL PT reagent Dade
Thromboplastin IS (Dade Diagnostics AG) was
added. The activities of the clotting factors VIII
and IX were determined by mixing 50 mL of a 1:10,
1:20, and 1:40 plasma dilution (Owren’s diluent
buffer) with 50 mL factor VIII- or factor IX-defi-
cient plasma (both from Immuno AG, Vienna,
Austria). The mixtures were then incubated for 60
seconds at 378C and thereafter 50 mL aPTT reagent
Dade Actin-FS (Dade Diagnostics AG) were
added; after a further incubation of 180 seconds at
378C, coagulation was started by the addition of
50 mL of 0.025 M CaCl2 (on the STA analyzer,
patient plasma, deficient plasma, and aPTT reagent
were incubated together for 240 seconds before
the addition of CaCl2). For the calibration of single
factor activity determination, dilutions of Coag Cal
N [Dade Diagnostics AG; 1:1 (5100%), 1:2, 1:4,
1:8, 1:16, and 1:32 with Owren’s diluent buffer]
with the respective factor-deficient plasma were
used. All dilutions of samples and plasma for cali-
bration were done automatically by the analyzers.
All coagulation tests measured on SYSMEX CA
6000 were performed by using a percent detection
method for coagulation end point determination
where the amount of light scattered immediately
after addition of reagent to the sample is taken as
0% and the amount of light scattered at comple-
tion of the reaction is taken as 100%. Clotting time
was determined at the 50% point. AT was de-
termined photometrically by a synthetic substrate
method, using serial measurements of changes
in absorbance at 405-nm wavelength. The STA
analyzer uses mechanical clot detection for coagu-
lation assays. The swinging of a small steel ball is
67
monitored because of alternating magnetic field.
Changes in the swing due to fibrin-clot formation
are taken as coagulation end point. For AT mea-
surement a chromogenic test (measurement at 405
nm) was used. Determinations of samples used for
the calibration of various tests were performed in
duplicate, whereas all the other measurements were
done in single tests. Fibrin split product D-dimer
was measured by ELISA (Asserachrom D-dimer,
Stago Diagnostics; normal range: ,400 mg/L).
For estimation of practicability in a routine coag-
ulation laboratory PT, aPTT, TT, fibrinogen, and
AT measurements in 40 individuals with values
within the normal range were performed. In these
samples a throughput of 123 analysis per hour
was calculated.
1.4. Statistical Analysis
For calculating coefficients of variation and regres-
sion analysis, after Passing/Bablok the statistic pro-
gram EVAPAK (Roche Austria) was used [3].
2. Results
2.1. Intraassay Precision
The intraassay precision of PT, aPTT, fibrino-
gen, TT, and AT were determined in pooled sam-
ples with values within the normal range and in
pathological samples (PT: liver disease, oral antico-
agulant treatment; aPTT: heparin treatment; TT:
heparin treatment; fibrinogen: hypo- and hyperfi-
brinogenemia; AT: liver disease). Values were ob-
tained from pooled plasma that was divided into
20 separate closed vacutainer tubes. The procedure
was repeated on the following two days. The values
given in Table 1 represent the mean values of the
results obtained on these 3 days. Intraassay precision
showed coefficients of variation (CVs) ,5% for
all tests within the normal range. In pathological
samples CVs were also ,5%, except for TT in
heparinized samples (5.4%), for fibrinogen in pa-
tients with hypofibrinogenemia (6.36%), and with
fibrinogen levels greater than 7.0 g/L (10.1.%), and
for decreased AT due to liver insufficiency
(9.26%).
Plasma pools of 10 healthy persons and of factor-
deficient patients were used to determine the in-
68 P. Quehenberger et al./Thrombosis Research 96 (1999) 65–71

Table 1. Intraassay precision in normal and abnormal plasma samples (n560)

PT (ranges) Healthy controls Liver disease patients OA-treated patients
(11.8–12.2 seconds), (15.0–22.3 seconds), (19.0–20.6 seconds),
0.65%60.09 0.64%60.11 0.67%60.13
aPTT (ranges) Healthy controls Heparin-treated patients
(30–41 seconds), (45–79 seconds),
1.47%60.45 4.42%62.57
TT (ranges) Healthy controls Heparin-treated patients
(14–20 seconds), (29–52 seconds),
3.47%60.64 5.40%62.05
Fib (ranges) Healthy controls Hypofibrinogenemia Hyperfibrinogenemia
(2.0–3.5 g/L), (,2.0 g/L), (3.5–7.0 g/L), 2.75%60.11,
4.37%60.13 6.36%60.08 (.7.0 g/L), 10.1%60.74
AT (ranges) Healthy controls Liver disease patients
(80–100%), (10–41%),
1.96%62.55 9.26%62.31
Percent CV6SD of three determinations are given.

traassay precision of coagulation factor measure-
ment. Each sample was divided into 20 separate
tubes. The mean values of the results calculated
from three dilutions (1:10, 1:20, 1:40) were used
for calculating intraassay precision; CVs were ,5%
for all tests in the normal and in the pathological
range (Table 2).
2.2. Interassay Precision and Carryover
samples precision values were ,5% for all factors
except for factor VII (6.12%) (Table 4).
To determine the amount of carryover of hepa-
rin, we analyzed the aPTT of pooled normal plasma
run after a sample of pooled normal plasma spiked
with three units of unfractionated heparin per mL.
In 10 alternating determinations of these two sam-
ples no prolongation of the aPTT of normal plasma
could be detected.

The interassay precision for PT, aPTT, TT, fibrino-
gen, and AT was tested over a period of 10 days
using aliquots of pooled and once-frozen plasma
samples. As shown in Table 3 the interassay preci-
sion within the normal range was below 5% for all
tests except for fibrinogen (5.62%). Precision for
abnormal plasmas was higher than 5% for aPTT
(11.6%), fibrinogen (11.7%), and AT (10.7%).
For calculating the interassay precision of single
factor activity determinations, values obtained on
10 days from frozen aliquots of patients plasma
samples were used. In normal and in pathological
2.3. Correlation between
Photo-Optical and Mechanical Clot Detection:
Basal Coagulation Tests, Single Factors,
and Optical Abnormal Plasma Samples
Coefficients of correlation between values ob-
tained with SYSMEX CA 6000 and with STA ana-
lyzer were calculated for PT, aPTT, TT, fibrinogen,
and AT in optical normal samples (r values be-
tween 0.96 and 0.992; n5100) (Table 5). Fibrinogen
levels measured with SYSMEX CA 6000 photo-
optical clot detection in 15 samples with markedly

Table 2. Intraassay precision of coagulation factors (n520)

Normal plasma samples Abnormal plasma samples
Mean SD CV (%) Mean SD CV (%)

FV (%) 88.1 4.3 4.85 26.0 0.9 3.50

FVII (%) 103.2 4.3 4.20 37.4 1.5 4.01
FVIII (%) 101.3 2.1 2.10 22.2 1.1 4.95
FIX (%) 82.3 1.4 1.76 26.2 1.3 4.96
 P. Quehenberger et al./Thrombosis Research 96 (1999) 65–71 69

Table 3. Interassay precision in commercial in normal and abnormal plasma samples

(n510)
Normal plasma samples Abnormal plasma samples
Mean SD CV (%) Mean SD CV (%)

PT (seconds) 12.8 0.2 1.6 20.3 0.9 4.4

APTT (seconds) 35.2 0.73 2.1 75.0 8.7 11.6
TT (seconds) 18.7 0.8 4.3 26.3 1.0 3.8
Fib (g/L) 2.67 0.15 5.62 1.62 0.19 11.7
AT (%) 90.2 4.4 4.9 33.7 3.6 10.7

elevated fibrin split product D-dimer levels (4800– 3. Discussion

410000 mg/L; mean 54075 mg/L) were compared
with those measured with STA mechanical clot
detection in order to assess the influence of fibrin
derivatives on clot detection. A good correlation
of fibrinogen values measured by the two analyzers
in samples with high fibrin split products was found:
r50.994, y50.872x10.120, p,0.0001. The activities
of clotting factors V (n514, r50.989, p,0.0001),
VII (n513, r50.993, p,0.0001), VIII (n512,
r50.996, p,0.0001), and IX (n513, r50.977,
p,0.0001) determined by SYSMEX CA 6000
showed good correlations with the values measured
by STA analyzer (Figure 1).
In order to assess the influence of common opti-
cal interferents on coagulation measurement, val-
ues determined by SYSMEX CA 6000 photo-opti-
cal clot detection in lipemic (triglyceride levels
were between 2.42–18.70 mmol/L), icteric (total
bilirubin levels were between 46–925 mmol/L), and
hemolytic plasma samples were compared with the
values from the same samples measured with me-
chanical clot detection using STA analyzer (n520
for each parameter). With the exception of TT,
adequate correlation between r50.902 and r50.995
were found between photo-optical and mechanical
clot detection in these samples (Table 5).
In the present study the coagulation analyzer SYS-
MEX CA 6000, which uses photo-optical clot de-
tection and is equipped with cap-piercing technique
and software for coagulation single factor deter-
mination, was evaluated with respect to its tech-
nical characteristics. Concerning intra- and in-
terassay precision, CVs were below 5% in most
assays evaluated. Clearly higher CVs were only
detected in abnormal samples [i.e., intraassay CV
of antithrombin measurement at low levels in liver
disease patients (CV59.26%), and fibrinogen mea-
surement in samples with fibrinogen levels below
2.0 g/L (CV56.36%) and greater than 7.0 g/L
(CV5 10.1%) or interassay CV of aPTT (11.6%),
AT (10.7%), fibrinogen (11.7%), and factor VII
(6.12%)]. The analyzer is provided with an auto-
matic cap-piercing system, which showed an excel-
lent mechanical performance. During several thou-
sands of determinations no defect of this system
was observed. The calculated CVs also demon-
strate that there are no influences of this cap-pierc-
ing system on the performance of the coagulation
tests. Automated cap piercing by SYSMEX CA
6000 saves technician time and also minimizes the
infectious risk for technicians, which is a big advan-
tage of this analyzer.

Table 4. Interassay precision of coagulation factors (n510)

Normal plasma samples Abnormal plasma samples
Mean SD CV (%) Mean SD CV (%)

FV (%) 90.6 4.0 4.4 29.3 1.4 4.8

FVII (%) 100.4 5.0 4.98 47.4 2.9 6.1
FVIII (%) 85.3 2.2 2.6 20.8 0.8 3.9
FIX (%) 80.1 1.6 2.0 26.0 0.7 2.7
70 P. Quehenberger et al./Thrombosis Research 96 (1999) 65–71

Table 5. Correlation coefficients of clotting assays performed on CA

6000 and STA analyzer (n5100)
Optically normal Lipemic Icteric Hemolytic

PT 0.975 0.907 0.974 0.987

APTT 0.960 0.988 0.971 0.905
TT 0.983 0.288 0.641 0.666
Fib 0.990 0.961 0.986 0.979
AT 0.992 0.902 0.985 0.995

The determination of clotting single factor activi-
ties are usually performed by laborious techniques
on semiautomated analyzers. Due to several manu-
ally performed pipetting and dilution steps these
determinations are of limited precision. In contrast,
single factor determinations on SYSMEX CA 6000
were found to be of satisfying precision. The possi-
bility of automatically performing determinations
on several dilutions of a plasma sample, as it is
executed on SYSMEX CA 6000, is important for
the exact interpretation of the measured values;
discrepancies between the results of several dilu-
tions may point to the existence of an inhibitor (e.g.,
a lupus anticoagulant). Therefore, the technician can
choose between the possibility of automatic averag-
ing or obtaining the values for each of the dilutions
separately shown. Mechanical clot detection (STA)
and photo-optical clot detection (SYSMEX CA
6000) showed good correlation concerning coagula-
tion single factor activity determinations.

Fig. 1. Correlation of clotting factor

activities determined on CA 6000 and
STA analyzer. (a) Correlation of clot-
ting factor V activities: y50.83x13.72,
r50.989, p,0.0001. (b) Correlation
of clotting factor VII activities: y5
1.04x16.52, r50.993, p,0.0001. (c)
Correlation of clotting factor VIII ac-
tivities: y51.08x20.15, r50.996, p,
0.0001. (d) Correlation of clotting fac-
tor IX activities: y51.09x15.13, r5
0.977, p,0.0001.
 P. Quehenberger et al./Thrombosis Research 96 (1999) 65–71
Photo-optical and mechanical clot detection also
showed satisfactory correlation for the basal coagu-
lation tests prothrombin time, activated partial
thromboplastin time, fibrinogen, and antithrom-
bin, not only in samples with normal optical quality
but also in optically abnormal specimens (e.g., in
lipemia, hyperbilirubinemia, and hemolysis). We
could also show that photo-optical and mechanical
clot detection do not show discrepancies in fibrino-
gen measurement using the method described by
Clauss [4], even in samples with markedly elevated
fibrin split products. The only test that showed a
weak correlation between photo-optical and me-
chanical clot detection was TT. The effect of the
interferents on TT measurement on SYSMEX CA
6000 may be due to the fact that plasma dilution
within this assay is less pronounced compared to
the other coagulation tests. We found a throughput
of 123 analyses per hour in normal samples using
SYSMEX CA 6000. This relatively low value can
be increased if the analyzer is integrated into a fully
automated “advanced coagulation analysis system”
offered by SYSMEX, consisting of an automated
centrifugation unit and several CA 6000 units.
Our data show that the automated coagulation
analyzer CA 6000 is a precise and reliable analyzer
71
for coagulation laboratories, not only with respect
to basal coagulation testing but also concerning the
determination of single factor clotting activities.
The comparison between values obtained with
photo-optical and with mechanical clot detection
showed that determinations on SYSMEX CA 6000
are rather insensitive to interferents. The cap-pierc-
ing technique, a new technique in coagulation test-
ing, is of adequate technical quality.
References
1. Kurosaki T. Overview of the Sysmex CA-6000
automated coagulation analyzer. Sysmex J Int
1996;6:51–5.
2. Kasai T, Motoi S. Principles of analysis by the
automated coagulation analyzer CA-6000. Sys-
mex J Int 1997;1:43–7.
3. Passing H, Bablok W. A new biometrical proce-
dure for testing the equality of measurement
from two different analytical methods. J Clin
Chem Clin Biochem 1983;21:709–20.
4. Clauss A. Gerinnungsphysiologische Schnell-
methode zur Bestimmung des Fibrinogens. Acta
Haemat 1957;17:237–46.